Evaluation of clustering results.
Class Summary Class Description BCubedBCubed measures for cluster evaluation. ClusterContingencyTableClass storing the contingency table and related data on two clusterings. EditDistanceEdit distance measures. EntropyEntropy based measures, implemented using natural logarithms. EvaluateClusteringEvaluate a clustering result by comparing it to an existing cluster label. EvaluateClustering.ParParameterization class. EvaluateClustering.ScoreResultResult object for outlier score judgements. LogClusterSizesThis class will log simple statistics on the clusters detected, such as the cluster sizes and the number of clusters. MaximumMatchingAccuracyCalculates the accuracy of a clustering based on the maximum set matching found by the Hungarian algorithm. PairCountingPair-counting measures, with support for "noise" clusters and self-pairing support. PairSetsIndexThe Pair Sets Index calculates an index based on the maximum matching of relative cluster sizes by the Hungarian algorithm. SetMatchingPuritySet matching purity measures.